100% Cpu usage constant due to /usr/bin/x

[westms]

An idea you can try. Not sure if it will work. When booting, in the LILO screen hit the TAB key. Now type "Salix acpi=off". See what happens. Also try "Salix acpi=force" and "Salix acpi_osi=".

Tried them all: 1) acpi=force, 2)acpi=off, and 3)acpi_osi= .

Hello,

if we want to continue to assume that there could be a problem with ACPI, please try also:

Code: Select all

Salix acpi_osi=Windows

[chazbroam]

Hello,

if we want to continue to assume that there could be a problem with ACPI, please try also:

Code: Select all

Salix acpi_osi=Windows

Here is my new lilo.conf append line:

Code: Select all

append="acpi_osi=\"!Microsoft Windows XP\" acpi=noirq acpi_pm_good pcie_aspm=on resume=/dev/sda2 vt.default_utf8=1"

Here is certain dmesg output:

Code: Select all

my_username[~]$ sudo dmesg | grep -i acpi
[    0.000000] BIOS-e820: [mem 0x000000003f7f0000-0x000000003f7f2fff] ACPI NVS
[    0.000000] BIOS-e820: [mem 0x000000003f7f3000-0x000000003f7fffff] ACPI data
[    0.000000] ACPI BIOS Bug: Error: A valid RSDP was not found (20130328/tbxfroot-218)
[    0.000000] Kernel command line: BOOT_IMAGE=Salix_gen ro root=801 acpi_osi="!Microsoft Windows XP" acpi=noirq acpi_pm_good pcie_aspm=on resume=/dev/sda2 vt.default_utf8=1
[    0.124510] PM: Registering ACPI NVS region [mem 0x3f7f0000-0x3f7f2fff] (12288 bytes)
[    0.276061] ACPI: Interpreter disabled.
[    0.277540] pci 0000:00:1f.0: quirk: [io  0x0400-0x047f] claimed by ICH4 ACPI/GPIO/TCO
[    0.281806] pnp: PnP ACPI: disabled
[   10.144553] ACPI Exception: AE_BAD_PARAMETER, Thread 4134187200 could not acquire Mutex [0x1] (20130328/utmutex-285)
[   10.150680] ACPI Exception: AE_BAD_PARAMETER, Thread 4134189280 could not acquire Mutex [0x1] (20130328/utmutex-285)
[   10.151549] ACPI Exception: AE_BAD_PARAMETER, Thread 4134189280 could not acquire Mutex [0x1] (20130328/utmutex-285)
[   10.154549] ACPI Exception: AE_BAD_PARAMETER, Thread 4134189280 could not acquire Mutex [0x1] (20130328/utmutex-285)

And the problem remains.

[westms]

Here is my new lilo.conf append line:

Code: Select all

append="acpi_osi=\"!Microsoft Windows XP\" acpi=noirq acpi_pm_good pcie_aspm=on resume=/dev/sda2 vt.default_utf8=1"

Here is certain dmesg output:

Code: Select all

my_username[~]$ sudo dmesg | grep -i acpi
[    0.000000] BIOS-e820: [mem 0x000000003f7f0000-0x000000003f7f2fff] ACPI NVS
[    0.000000] BIOS-e820: [mem 0x000000003f7f3000-0x000000003f7fffff] ACPI data
[    0.000000] ACPI BIOS Bug: Error: A valid RSDP was not found (20130328/tbxfroot-218)
[    0.000000] Kernel command line: BOOT_IMAGE=Salix_gen ro root=801 acpi_osi="!Microsoft Windows XP" acpi=noirq acpi_pm_good pcie_aspm=on resume=/dev/sda2 vt.default_utf8=1
[    0.124510] PM: Registering ACPI NVS region [mem 0x3f7f0000-0x3f7f2fff] (12288 bytes)
[    0.276061] ACPI: Interpreter disabled.
[    0.277540] pci 0000:00:1f.0: quirk: [io  0x0400-0x047f] claimed by ICH4 ACPI/GPIO/TCO
[    0.281806] pnp: PnP ACPI: disabled
[   10.144553] ACPI Exception: AE_BAD_PARAMETER, Thread 4134187200 could not acquire Mutex [0x1] (20130328/utmutex-285)
[   10.150680] ACPI Exception: AE_BAD_PARAMETER, Thread 4134189280 could not acquire Mutex [0x1] (20130328/utmutex-285)
[   10.151549] ACPI Exception: AE_BAD_PARAMETER, Thread 4134189280 could not acquire Mutex [0x1] (20130328/utmutex-285)
[   10.154549] ACPI Exception: AE_BAD_PARAMETER, Thread 4134189280 could not acquire Mutex [0x1] (20130328/utmutex-285)

And the problem remains.

Code: Select all

append="acpi_osi=\"!Microsoft Windows XP\" acpi=noirq acpi_pm_good pcie_aspm=on resume=/dev/sda2 vt.default_utf8=1"

What or who tells you that !Microsoft Windows XP is a string, which the BIOS can understand?

First, what does the exclamation mark in front of Microsoft? Usually it means NOT or NON. But it is not part of the name. What is the significance here?

Then, perhaps Microsoft Windows XP is a string that is known by the HP BIOS. But according to my knowledge, it is unknown. The Operating System Vendor String Prefix are by definition: FreeBSD, HP-UX, Linux, OpenVMS and Windows. Perhaps Windows 2001, Microsoft Windows NT, Microsoft Windows and Microsoft Windows ME are also permitted by the BIOS manufacturer.

Then to the dmesg extract. The error message:

Code: Select all

[    0.000000] ACPI BIOS Bug: Error: A valid RSDP was not found (20130328/tbxfroot-218)

should already be borne in mind and should not be ignored. The RSDP could not be created. Quite possible that the broken acpi_osi string is responsible.

The message:

Code: Select all

[    0.276061] ACPI: Interpreter disabled.

is noteworthy. Your ACPI system is not running.

Then there are four times this message:

Code: Select all

[   10.144553] ACPI Exception: AE_BAD_PARAMETER, Thread 4134187200 could not acquire Mutex [0x1] (20130328/utmutex-285)

One can not ignore such messages when something does not work then.

In addition:
acpi=noirq : Why?
acpi_pm_good pcie : Why?
pcie_aspm=on : Is that necessary at this time of testing?
Unfortunately, you do not explain why these parameters are set.

Further, it seems to me as if you had always reached runlevel 4, you logged in, and then changed to runlevel 3. But this is of evil.

If you want to find the error and still have enough energy, I suggest the following procedure:

First: Such a collection of parameters in the append line is very unfavorable for testing. The LILO append line is truncated to:

Code: Select all

append="pcie_aspm=on resume=/dev/sda2 vt.default_utf8=1"

Run LILO and then shutdown.

Second: Let the system run first only in runlevel 1 as follows:

1st: When the hardware is off, switch off power supply / pull power plug. Wait 10 seconds. Start with a cold boot, because there is a difference between cold boot and warm boot in the BIOS.

2nd: Switch power on and boot.

3rd: When the LILO screen is displayed, hit the TAB key.

4th: Enter in the LILO command line:

Code: Select all

Salix_gen 1

5th: Log in. Does the problem persist? If so, then shutdown as superuser / root with:

Code: Select all

shutdown -h now

For

Code: Select all

Salix_gen 1 acpi=force
Salix_gen 1 acpi=off
Salix_gen 1 acpi_osi=Windows
Salix_gen 1 acpi_osi=Linux
Salix_gen 1 acpi_osi=

repeat steps 1st, 2nd, 3rd and 5th.

Also run meaningful combinations thereof, e.g.:

Code: Select all

Salix_gen 1 acpi=force acpi_osi=Windows

If the problem does not occur in run level 1, then try the previous in runlevel 3:

Code: Select all

Salix_gen 3
Salix_gen 3 acpi=force
...
...
...

The configuration acpi=off is intended for installation only when all else fails. It should be avoided to have a so setup system running, because the cooling could not function and the CPU / chipset would burn up.

Incidentally: With Suse and Ubuntu, I had the same problem in which the shutdown did not come to an end and for example, the desktop background image remained. There are two HP Vectra VL400, now very old. Later the desktops were not fully established after each boot. These problems I could not solve. After a long period of suffering I had installed Zenwalk, which worked just fine.

[chazbroam]

Code: Select all

append="acpi_osi=\"!Microsoft Windows XP\" acpi=noirq acpi_pm_good pcie_aspm=on resume=/dev/sda2 vt.default_utf8=1"

What or who tells you that !Microsoft Windows XP is a string, which the BIOS can understand?

First, what does the exclamation mark in front of Microsoft? Usually it means NOT or NON. But it is not part of the name. What is the significance here?

Then, perhaps Microsoft Windows XP is a string that is known by the HP BIOS. But according to my knowledge, it is unknown. The Operating System Vendor String Prefix are by definition: FreeBSD, HP-UX, Linux, OpenVMS and Windows. Perhaps Windows 2001, Microsoft Windows NT, Microsoft Windows and Microsoft Windows ME are also permitted by the BIOS manufacturer.

I have seen previous documentation on acpi append statements and followed what I saw there. I think it may have been for end users of ARCH Linux, but it was for acpi itself, so I copied it to my lilo.conf append statement.

acpi=noirq : Why?

Because I was trying different things to see the response.

acpi_pm_good : Why?

Because when you put acpi=force in the append line, dmesg reports a line saying something about a bug and to use "acpi_pm_good" as some kind of workaround.

pcie_aspm=on

It fixed something else, so I kept it.

Further, it seems to me as if you had always reached runlevel 4, you logged in, and then changed to runlevel 3. But this is of evil.

Correction:

• 1) I originally logged in as runlevel 4 & logged in,
  2) then edited /etc/Inittab to run in runlevel 3, so I would not have to keep hitting the tab button and typing "Salix_gen init 3".
  3) this is not evil, it's intelligence in action.

First: The LILO append line is truncated to:

Code: Select all

append="pcie_aspm=on resume=/dev/sda2 vt.default_utf8=1"

Run LILO and then shutdown.

Done.

Second: Let the system run first only in runlevel 1 as follows:

1st: When the hardware is off, switch off power supply / pull power plug. Wait 10 seconds. Start with a cold boot, because there is a difference between cold boot and warm boot in the BIOS.

2nd: Switch power on and boot.

3rd: When the LILO screen is displayed, hit the TAB key.

4th: Enter in the LILO command line:

Code: Select all

Salix_gen 1

5th: Log in. Does the problem persist? If so, then shutdown as superuser / root with:

Code: Select all

shutdown -h now

For

Code: Select all

Salix_gen 1 acpi=force
Salix_gen 1 acpi=off
Salix_gen 1 acpi_osi=Windows
Salix_gen 1 acpi_osi=Linux
Salix_gen 1 acpi_osi=

repeat steps 1st, 2nd, 3rd and 5th.

Done. All steps give the same response (except for "Salix_gen 1 acpi=off", which I didn't try):
Htop gives now intermittent 100% CPU usage, like every 3 seconds the lines at the top will start out at 0%, then all light up as GREEN, but report 100%, then go RED and report 100%, then back to 0% or some other low, low number. And then it all starts all over in an endless loop.

ACPI DOES NOT EVER START -- EVER!!

Code: Select all

sudo dmesg | grep -i acpi

always says "ACPI: Interpreter disabled."

Also run meaningful combinations thereof, e.g.:

Code: Select all

Salix_gen 1 acpi=force acpi_osi=Windows

Didn't try this, since ACPI never comes on no matter what.

[westms]

[...] Htop gives now intermittent 100% CPU usage, like every 3 seconds the lines at the top will start out at 0%, then all light up as GREEN, but report 100%, then go RED and report 100%, then back to 0% or some other low, low number. And then it all starts all over in an endless loop.

ACPI DOES NOT EVER START -- EVER!!

Code: Select all

sudo dmesg | grep -i acpi

always says "ACPI: Interpreter disabled."

When examine the problem, what happened to error message: ACPI BIOS Bug: Error: A valid RSDP was not found (20130328/tbxfroot-218)?

Can ACPI be affected in the BIOS settings?

Are plug-in cards in the computer, which are not ACPI compatible, or which have been developed for ACPI 2.0 or older?

Do you have a manual for this HP hardware, where the ACPI capabilities are described? Can you get one?


The ACPI is running incorrectly, although other approaches:

The problem also exists in runlevel 1. In runlevel 1 the kernel modules are already loaded. Do you have kernel modules installed, you could be held responsible for this behavior?

Did you add program launches to scripts in /etc/rc.d, especially in rc.K and rc.S?

[rokytnji]

If I am off the mark. Just ignore my post. I have ran into newer xorg not playing nice with legacy modules for video in the kernel.
Mainly savage chips and sis chips. My cpu would hammer at 100% also.

My posts on this subject.
http://antix.freeforums.org/post40380.html#p40380

Then moving through that thread
http://antix.freeforums.org/post40383.html#p40383
http://antix.freeforums.org/post40385.html#p40385
http://antix.freeforums.org/post40401.html#p40401
http://antix.freeforums.org/post40421.html#p40421

Like I said. If I am off base here with recommending running vesa to bring down cpu hammering when in your desktop.
Then just pretend I am not here.

[chazbroam]

When examine the problem, what happened to error message: ACPI BIOS Bug: Error: A valid RSDP was not found (20130328/tbxfroot-218)?

Nothing, as far as I know, it's still there.

Can ACPI be affected in the BIOS settings?

No. I just flashed the bios to the most updated BIOS for this older 32bit HP computer. No BIOS settings for ACPI or APM.

Are plug-in cards in the computer, which are not ACPI compatible, or which have been developed for ACPI 2.0 or older?

I don't know. It started out with an Intel 82845G Brookdale-G onboard, so you tell me. My guess -- probably. BUT, I added a new Radeon HD 3450 graphics card, and have been using it instead. This newer card uses the available PCI add-on port the computer has on it's motherboard (all other GPU cards needed PCIe).

Do you have a manual for this HP hardware, where the ACPI capabilities are described? Can you get one?

No, and no.

The problem also exists in runlevel 1. In runlevel 1 the kernel modules are already loaded. Do you have kernel modules installed, you could be held responsible for this behavior?

Haven't added any kernel modules, but I think the Radeon GPU needs the kernel modesetting on, which IT DOES by default, NOT ME. STOP BLAMING ME!!!

Did you add program launches to scripts in /etc/rc.d, especially in rc.K and rc.S?

No, haven't touched those scripts at all, not even rc.local.

Now, I was able to get ACPI started after initial booting. I just added a line to rc.local -- just now -- and it reported in the Xorg.0.log that ACPI is running finally.

But, the problem remains.